Hi, I'm Diego

Full-Stack AI Engineer

Meta Llama Impact Grant Winner · $100K

I build production AI systems — agents, RAG, and edge LLMs — end to end.

About.

I'm a product-focused Full-Stack AI Engineer specializing in agentic workflows (LangGraph) and high-performance backends (FastAPI). I ship production RAG, fine-tuning, and edge AI systems for startups and enterprise.

My path was self-taught. I started in bilingual customer service, moved up to the escalations team at Teleperformance, then left to study programming full-time — which gave me the communication, English fluency, and grit I bring to engineering today.

Now I co-found and build AI products that reach real users: from cloud reasoning with Llama and tool-calling, to quantized edge models that run fully offline when the network goes down.

Mexico

Projects.

Featured

BluAI — AI Hurricane Resilience Platform

$100K Meta Llama Impact Grant winner. Hybrid cloud/edge AI hurricane assistant.

BluAI is an AI-powered hurricane resilience platform and the core deliverable of our $100K Meta Llama Impact Grant. I architected a hybrid AI system — Llama 3.3 in the cloud (RAG + tool-calling) for complex reasoning, switching to quantized Llama edge models (PyTorch ExecuTorch / PTE) for fully offline inference during connectivity blackouts — and built the full-stack mobile app in React Native + FastAPI with map-based risk visualization and personalized disaster alerts. Now live on Google Play, with iOS in active development.

React NativeFastAPIPythonLlamaRAGPyTorch

Google Play Website Code

Featured

AI Spine — Autonomous Agent Platform

Spin up autonomous AI agents from natural language. Heavy LangChain + LangGraph.

AI Spine is a platform for building autonomous agents using natural language. I engineered the core architecture and implemented complex agentic memory and tool-calling workflows with LangChain and LangGraph, enabling agents to retain context across sessions and platforms. Delivered a modular, production-ready MVP in FastAPI/Python with a full technical handover; retained as technical advisor. (Contract project for Dataframe — internals private.)

PythonFastAPILangChainLangGraph

Website

Featured

The Overseer — LangGraph Productivity Agent

A vigilante agent that enforces deadlines and escalates via WhatsApp & calls.

A personal LangGraph agent that enforces task deadlines by polling a source of truth (Notion API), then escalates interventions through Twilio (WhatsApp → phone calls) as deadlines approach. Backend built with FastAPI, PostgreSQL (SQLAlchemy), and APScheduler for persistent task state and scheduled polling. A practical demo of agentic state-machine patterns, scheduling, and human-in-the-loop escalation.

LangGraphFastAPIPythonPostgreSQLTwilioNotion

Featured

Harris & Frank — AI Suit Visualizer

Photorealistic suit previews generated from raw fabric textures — no photoshoot needed.

A production web app for sales teams that generates photorealistic suit previews from raw fabric textures, replacing physical photoshoots. I fine-tuned custom SDXL + LoRA models on RunPod, with a full-stack architecture in Next.js + FastAPI + Cloudflare R2. (Client project — internals private.)

PythonFastAPINext.jsLoRAPyTorchRunPodCloudflare

Experience.

Co-Founder & Lead AI Engineer · BluAI

2024 — Present

Winner — Meta Llama Impact Grant ($100K USD)

Co-founded and lead technical development for BluAI, a hurricane resilience platform launching to production on Google Play in May 2026, with iOS in active development.
Architected a hybrid AI system: Llama 3.3 (cloud) with RAG and tool-calling for complex reasoning, switching to quantized Llama edge models (PTE) for fully offline inference during connectivity blackouts.
Engineered the full-stack mobile app in React Native + FastAPI, integrating map-based risk visualization, personalized disaster alerts, and a hybrid cloud/edge inference pipeline.

React NativeFastAPIPythonLlamaRAGRailway

Full-Stack AI Engineer · Dataframe — AI Spine

2025

Engineered the core architecture for AI SPINE, a platform for building autonomous agents using natural language.
Implemented complex agentic memory and tool-calling workflows using LangChain and LangGraph, enabling agents to retain context across sessions and platforms.
Delivered a modular, production-ready MVP (FastAPI/Python) and executed a full technical handover to the founding team. Retained as technical advisor.

PythonFastAPILangChainLangGraphPostgreSQL

Full-Stack AI Engineer · Harris & Frank

2025

Built a production web app for sales teams that generates photorealistic suit previews from raw fabric textures, replacing physical photoshoots.
Fine-tuned custom SDXL + LoRA models on RunPod; full-stack architecture in Next.js + FastAPI + Cloudflare R2.

PythonFastAPINext.jsLoRARunPodCloudflare

Personal Project · The Overseer

2026 — Present

Built a personal LangGraph agent that enforces task deadlines by polling a source of truth (Notion API), then escalates interventions via Twilio (WhatsApp → calls) as deadlines approach.
Engineered the backend with FastAPI, PostgreSQL (SQLAlchemy), and APScheduler for persistent task state and scheduled polling.

LangGraphFastAPIPythonPostgreSQLTwilioNotion

Co-Founder · Verskod

2024 — 2026

Co-founded a development consultancy delivering full-stack and AI engineering for early-stage startups.
Led technical scoping and Web + FastAPI + PostgreSQL builds. Exited as co-founder; retained as technical advisor.

PythonFastAPINext.jsPostgreSQL

Achievements.

Meta Llama Impact Grant Winner

Won the $100K Meta Llama Impact Grant with BluAI, an AI-powered hurricane resilience system recognized by Meta for social impact.

2024 — 2025

Featured at LlamaCon 2025

Showcased BluAI as one of Meta's top AI impact projects at their global LlamaCon event.

2025

TalentLand 2025 — Google Cloud Scholarship

Attended on full scholarship. Joined AI, dev, and startup-focused workshops and networking events.

2025

NASA Space Apps Challenge — CDMX

Took part in NASA's global hackathon focused on solving Earth and space challenges using open data. Co-created 'MetaQuetzal'.

2024

Speaker at PROYCOMS

Spoke at an online conference on stress, work-life balance, and workplace culture.

2024

Skills.

AI & ML

Agentic AI (LangGraph, LangChain)
RAG Pipelines
LLM Fine-tuning (LoRA)
Edge LLM Deployment (ExecuTorch / PTE)
RunPod
OpenAI API

Backend

Python
FastAPI
Pydantic
PostgreSQL
Redis
Celery

Frontend

Next.js
React Native
Tailwind CSS
TypeScript

Deployment

Vercel
Railway
Docker
Git
GitHub
CI/CD Workflows

Languages

English (C1)
Spanish (Native)

Contact.

Open to roles and collaborations building production AI — agents, RAG, and backends. The fastest way to reach me is email.

luiscolin764@gmail.com LinkedIn GitHub